Itinera

Anisotropic least-cost path (LCP), corridor (LCC), From-Everywhere-To-Everywhere (FETE), stochastic / probabilistic paths and route validation (Path Deviation Index + buffer overlap) primitives for movement modelling — a pure numpy / scipy library extracted from the Itinera QGIS plugin.

Cost is directional (uphill ≠ downhill): each DEM cell becomes a graph node, edge weights come from a directional cost function of the signed slope, so the conductance matrix is asymmetric — true anisotropy. Paths are solved with scipy.sparse.csgraph.dijkstra.

pip install itinera

Only numpy and scipy are required.

Quick start

import numpy as np
from itinera import build_conductance, least_cost_path, tobler, rowcol_to_node

# A small DEM (elevations in metres) on a 10 m grid.
dem = np.random.default_rng(0).random((50, 50)) * 100.0

matrix, rows, cols = build_conductance(
    dem, cellsize=10.0, cost_fn=tobler, neighbours=8)

origin = rowcol_to_node(0, 0, cols)
dest = rowcol_to_node(49, 49, cols)
path, total_cost = least_cost_path(matrix, origin, dest)  # path = node indices

Turn node indices back into (row, col) with node_to_rowcol(node, cols).

What's included

• Cost functions (ten): tobler, tobler_offpath, herzog, naismith, llobera_sluckin, irmischer_clarke, minetti, pandolf, wheeled, pack_animal — each (slope, distance, **params) -> cost. pandolf reads mass / load / terrain; wheeled / pack_animal are experimental heuristics — anisotropic critical-slope presets with illustrative default thresholds, not calibrated published functions.
• Conductance: build_conductance (slope, optional barrier/multiplier, cost_params), build_conductance_friction (friction raster, optional DEM).
• Paths: accumulated_cost, least_cost_path, corridor / corridor_band, fete.
• Accessibility: accessibility — cost-distance surface + catchment mask + isochrone bands from source(s).
• Randomized shortest paths: rsp_passages — a θ-tunable movement-density surface that spans the least-cost path (θ→∞) and the random-walk / circuit current (θ→0), plus the RSP free-energy distance.
• Circuit theory: current_density (graph-Laplacian current + pinch points) and restoration_score (McRae 2012 barrier / restoration improvement map).
• Multi-criteria: composite_friction — a generic heuristic combiner (not a Litvine 2024 reproduction) merging several penalty rasters into one friction multiplier (weighted sum/product, per-layer invert, NoData masks); min-max normalisation makes it extent-dependent.
• Stochastic: stochastic_lcp (DEM-error + edge-drop + cost-model randomisation via cost_fns / cost_weights / param_jitter, with an early-stop convergence criterion via tol / convergence / return_diagnostics), add_dem_error, simulate_error_field (variogram-based DEM error field), add_global_stochasticity.
• Validation: pdi (Jan et al. 2000: area / straight-line O–D distance, endpoints snapped to the reference O/D), buffer_overlap (Goodchild & Hunter buffer method), mean_pairwise_overlap (route-stability across a set of paths).
• Grid helpers: xy_to_rowcol, check_/assert_regular_geotransform, check_/assert_grids_aligned.
• Utilities: estimate_conductance_bytes, format_bytes, block_reduce_mean.

Scope: bring your own raster I/O

This is a numerics library — it works on numpy arrays + a GDAL-style geotransform tuple. It does not depend on GDAL and does not read or write raster files. Load your DEM with whatever you already use (rasterio, osgeo.gdal, xarray/rioxarray, …) and pass the array in.

Use a projected CRS in metres so slope and distance are metric.

Relation to the QGIS plugin

The same code ships inside the Itinera QGIS plugin (where it is the plugin's private core package, plus QGIS/GDAL wrappers). The PyPI package and the QGIS plugin share the top-level import name itinera; they are meant for separate environments. Don't pip install itinera into the Python interpreter that runs QGIS — the plugin already bundles this code, and the two would shadow each other on sys.path.

Licence

MIT — see LICENSE. The QGIS plugin wrapper is licensed separately under GPLv3. References for the methods (Tobler, Naismith, Herzog, Llobera & Sluckin, Irmischer & Clarke, Minetti, Pandolf/Santee, White & Barber, Lewis, Jan et al. for PDI, Goodchild & Hunter for buffer overlap, Panzacchi/Saerens & van Etten for RSP, McRae for circuit theory) are in docs/REFERENCES.md.